Academic journal article Journal of Business and Behavior Sciences

Behavior of Monthly Total Returns of U.S. Treasury Bills: 1926 to 2011

Academic journal article Journal of Business and Behavior Sciences

Behavior of Monthly Total Returns of U.S. Treasury Bills: 1926 to 2011

Article excerpt


The U.S. Treasury bills market is one of the most active markets in the world. U.S. Treasury securities are the debt financing instruments of the United States federal government. T-bills have maturities of 1-month, 3-month, 6-month, and 1-year. They are issued at a discount to face value and reach full value (par value) at maturity. The difference between the discount price and par value is the interest or return earned. They carry a par value of $1,000 and up to $5 million. T-bills are backed by the credit of the U.S. government, and are thus considered close to riskfree investment. From 1928 to 2011, U.S. three-month T-bills provided investors with a 3.61% compound annual return. With 3.11% compound annual inflation, the real return over this period has been 0.50%. During this period, the annual returns have ranged between 0% and 15%, with the returns ranging from 0% to 5% in 59 of the 84-year period. Historically, T-bills have realized the highest returns during the inflationary 1970's and early 1980's. T-bills have realized the lowest returns during the years of the Great Depression (the 1930's), during the 1940's, when interest rates were price controlled ("pegged") by the Federal Reserve, and recently during the "Great Recession" period and thereafter (2008 - 2011).

With such an active trading in T-bills and with sophisticated investors that include many foreign governments, the T-bills market is expected to exhibit a high degree of efficiency. For example, we will not expect to find monthly seasonality in the returns from T-bills. This paper explores some aspects of the behavior of the monthly returns from T-bills in general and the presence of monthly seasonality in particular. Insight into the behavior of U.S. T-bills returns has implications for investors and policy makers. The next section deals with previous research on the U.S. T-bills returns behavior, followed by research methodology of this study, analysis of results, and summary and conclusion.


Contrasted to the extensive research on equity returns, few investigations examine seasonality in the fixed income markets and even fewer in the T-bills market. Gibbons and Hess (1981) found some day-of-the-week effects in the trading of thirty-day T-bills between 1963 and 1968. Ferri,oldstein, and Oberhelman (1984) also look at the day-of-the-week effect in T-bills return. The result of this study show for the data tested that the day-of-the-week effect in the returns of the bill market are occasional but irregular features of the market. Eiseman and Timme (1984) explore intraweek seasonality in the federal funds market using data from January 1966 to June 1982, and found seasonality to vary over time in intensity and relative size. Park and Reinganum (1986) find unusual price behavior of Tbills that mature at the turn of calendar months. Flannery and Protopapadakis (1988) find intrawcek seasonality continue to be significant but not uniform. Chen and Chan (1997) found using monthly returns from 1926 to 1990 T-bills return showed strong October effect during economic expansion and strong November effect during contraction.


The goal of this research is to find if there was a month effect in U.S.T-bills total monthly returns for the period of the study (1926-2011) and if so, was it more pronounced during Democratic presidencies or Republican presidencies. Many studies have used the dummy variable methodology to detect market seasonality. Chien, Lee and Wang (2002) provide statistical analysis and empirical evidence that the methodology may lead to misleading results. We avoid this problem by following the methodology used in Hamid and Dhakar (2005) through which they analyze seasonality in the monthly changes of the Dow Jones Industrial Average.

We study the month effect in three different ways. Unless otherwise stated, significance in all cases is tested at the 5% level.

1. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.